Tech companies including Cadence have done their share of acquiring startups to access new technologies and products. Most recently Cadence bid to buy Tensilica in Silicon Valley and Cosmic Circuits in Bangalore to expand its portfolio of intellectual property cores.
Cadence has plenty in the bank for any other acquisition candidates it finds. The company has $827 million in cash, generates more cash each quarter and has a $250 million line of credit at a relatively low interest rate. “We generate a lot of cash, so liquidity is not our barrier,” said Tan.
The chip designers the EDA sector serves, however, face a tougher financial outlook. Market watcher Semico Research predicts the $108 million required to design a 28-nm chip could balloon to $210 million at the 14-nm node.
The costs will drive more consolidation in big mobile and server platforms, Tan said. Meanwhile, analog, automotive and industrial parts will avoid the bleeding edge processes, he said.
Those dynamics could dampen growth for the EDA sector. That’s one reason why Cadence is joining Synopsys in expanding its offerings of IP cores, a business Semico says is growing about 19 percent a year.
“When you show double-digit growth, people start to invest in you and when you are not growing they ask for a dividend,” said Tan, complaining the EDA sector is severely undervalued today at its current 2.4-2.5 multiples. By contrast, service companies such as Salesforce.com trade at eight times their earnings, he added.
A full service of integrated cores and tools can also help systems companies innovate faster, expanding the entire electronics pie for everyone, he believes. “If the electronics industry grows from $300 billion to $500 billion, we are all better off--otherwise we eat each other’s lunch,” he said.
EDA companies are all in the same boat. “If one stock goes down, we all go down,” he said.
“We should not let people think EDA is only a $5 to $6 billion industry,” Tan said. “We need discipline about articulating our value proposition…growing from EDA to IP to [platform] software and services,” he said.
Tan added one more item to the EDA to-do list—be more fun. “We need more excitement so that rather than join Facebook or Google we can get college kids to join EDA companies,” he said.
"and its painful to see your kids go to the other side of the world" don't get it.. talking like a priest there, what about people who went the other way.. nothing wrong in going to India or China to work..
“but I would hate to see those countries have the most semiconductor companies." HATE, why??
Lip Bu Tan: US is not the start and end of the world. They are the biggest market, don't make it sound like a war cry just yet.
"There's money to be made in chip startups," the big EDA companies are usually not much interested in startups where there isn't much money to be made. Atleast this is my experience in dealing with a major EDA company on behalf of a startup, several years back.
The major cost barrier for any semi startup is the EDA tool cost. Maybe the big EDA companies should do something to address this issue first. Something like give away the tools at very low cost and collect back royalties, when any money is made.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.